Re: [PATCH v1] powerpc: Make setjmp/longjump signature standard

From: Nathan Chancellor <natechancellor@gmail.com>
To: Christophe Leroy <christophe.leroy@c-s.fr>
Cc: Clement Courbet <courbet@google.com>,
	Nick Desaulniers <ndesaulniers@google.com>,
	Michael Ellerman <mpe@ellerman.id.au>,
	Benjamin Herrenschmidt <benh@kernel.crashing.org>,
	Paul Mackerras <paulus@samba.org>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	Segher Boessenkool <segher@kernel.crashing.org>,
	linuxppc-dev@lists.ozlabs.org, linux-kernel@vger.kernel.org,
	clang-built-linux@googlegroups.com
Subject: Re: [PATCH v1] powerpc: Make setjmp/longjump signature standard
Date: Fri, 27 Mar 2020 11:27:08 -0700	[thread overview]
Message-ID: <20200327182708.GA13308@ubuntu-m2-xlarge-x86> (raw)
In-Reply-To: <f1b85a2a-1c60-9a12-f547-13ff255f18f0@c-s.fr>

On Fri, Mar 27, 2020 at 06:45:21PM +0100, Christophe Leroy wrote:
> Subject line, change longjump to longjmp
> 
> Le 27/03/2020 à 11:07, Clement Courbet a écrit :
> > Declaring setjmp()/longjmp() as taking longs makes the signature
> > non-standard, and makes clang complain. In the past, this has been
> > worked around by adding -ffreestanding to the compile flags.
> > 
> > The implementation looks like it only ever propagates the value
> > (in longjmp) or sets it to 1 (in setjmp), and we only call longjmp
> > with integer parameters.
> > 
> > This allows removing -ffreestanding from the compilation flags.
> > 
> > Context:
> > https://lore.kernel.org/patchwork/patch/1214060
> > https://lore.kernel.org/patchwork/patch/1216174
> > 
> > Signed-off-by: Clement Courbet <courbet@google.com>
> > ---
> >   arch/powerpc/include/asm/setjmp.h | 6 ++++--
> >   arch/powerpc/kexec/Makefile       | 3 ---
> >   2 files changed, 4 insertions(+), 5 deletions(-)
> > 
> > diff --git a/arch/powerpc/include/asm/setjmp.h b/arch/powerpc/include/asm/setjmp.h
> > index e9f81bb3f83b..84bb0d140d59 100644
> > --- a/arch/powerpc/include/asm/setjmp.h
> > +++ b/arch/powerpc/include/asm/setjmp.h
> > @@ -7,7 +7,9 @@
> >   #define JMP_BUF_LEN    23
> > -extern long setjmp(long *) __attribute__((returns_twice));
> > -extern void longjmp(long *, long) __attribute__((noreturn));
> > +typedef long *jmp_buf;
> 
> Do we need that new opaque typedef ? Why not just keep long * ?

Yes, otherwise the warning comes back:

In file included from arch/powerpc/kexec/crash.c:25:
arch/powerpc/include/asm/setjmp.h:10:12: error: declaration of built-in function 'setjmp' requires the declaration of the 'jmp_buf' type, commonly provided in the header <setjmp.h>. [-Werror,-Wincomplete-setjmp-declaration]
extern int setjmp(long *env) __attribute__((returns_twice));
           ^
arch/powerpc/include/asm/setjmp.h:11:13: error: declaration of built-in function 'longjmp' requires the declaration of the 'jmp_buf' type, commonly provided in the header <setjmp.h>. [-Werror,-Wincomplete-setjmp-declaration]
extern void longjmp(long *env, int val) __attribute__((noreturn));
            ^
2 errors generated.

> > +
> > +extern int setjmp(jmp_buf env) __attribute__((returns_twice));
> > +extern void longjmp(jmp_buf env, int val) __attribute__((noreturn));
> >   #endif /* _ASM_POWERPC_SETJMP_H */
> > diff --git a/arch/powerpc/kexec/Makefile b/arch/powerpc/kexec/Makefile
> > index 378f6108a414..86380c69f5ce 100644
> > --- a/arch/powerpc/kexec/Makefile
> > +++ b/arch/powerpc/kexec/Makefile
> > @@ -3,9 +3,6 @@
> >   # Makefile for the linux kernel.
> >   #
> > -# Avoid clang warnings around longjmp/setjmp declarations
> > -CFLAGS_crash.o += -ffreestanding
> > -
> >   obj-y				+= core.o crash.o core_$(BITS).o
> >   obj-$(CONFIG_PPC32)		+= relocate_32.o
> > 
> 
> Christophe